Definition Quality Scoring for Machine Learning Training Data

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

Existing data governance systems face challenges in creating and approving clear, unambiguous definitions for business assets, leading to contradictions and ambiguities in data interpretation and organization.

Innovation Solution

A system and method for automatically assessing the quality of definitions using machine learning models and guidelines, which evaluate structure, conciseness, circularity, and understandability, providing a score and allowing user feedback to improve definitions.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Manufacturing precision

If manual definition creation and approval processes are used, then definitions can be carefully crafted and reviewed, but the process is time-consuming and prone to human error and inconsistency

Engineering Contradiction:
Improvedefinition qualityVSAvoiddefinition creation time
Core Design Contradiction:
Manufacturing precisionVSLoss of time

Solution Approach 1:

The system performs preliminary automated evaluation of definitions against established quality guidelines before human review, pre-identifying issues with structure, conciseness, circularity, and understandability. This preliminary action filters out obviously defective definitions and prepares quality assessments in advance, reducing the time burden on manual reviewers while maintaining high quality standards.

Inventive Principle:
Principle #10Preliminary action

Solution Approach 2:

The patent replaces manual mechanical review processes with automated machine learning models that evaluate definitions against quality guidelines. The ML system automatically assesses structural correctness, conciseness, circularity detection, and understandability metrics, substituting human mechanical review with computational analysis that operates faster and more consistently.

Inventive Principle:
Principle #28Mechanics substitution (Replace mechanical system)

2Reliability

If multiple reviewers manually evaluate definitions, then comprehensive quality assessment is achieved, but resource consumption and coordination complexity increase

Engineering Contradiction:
Improvedefinition quality assessmentVSAvoidreview process complexity
Core Design Contradiction:
ReliabilityVSDevice complexity

Solution Approach 1:

The quality assessment process is segmented into distinct evaluation dimensions: structure guidelines, conciseness metrics, circularity detection, and understandability measures. Each dimension is evaluated by specialized machine learning models, allowing comprehensive assessment without requiring multiple human reviewers to coordinate all aspects. The segmentation enables parallel processing of different quality aspects.

Inventive Principle:
Principle #1Segmentation

Solution Approach 2:

The machine learning evaluation system acts as an intermediary between definition creators and human reviewers. It provides automated quality scores and feedback that guide manual review efforts, reducing the burden on human reviewers while maintaining comprehensive quality assessment. The intermediary system consolidates multiple evaluation criteria into unified feedback.

Inventive Principle:
Principle #24Intermediary (Mediator)

3Productivity

If automated evaluation systems are implemented, then processing speed increases, but accuracy and nuance in quality assessment may decrease

Engineering Contradiction:
Improvedefinition evaluation speedVSAvoidquality assessment accuracy
Core Design Contradiction:
ProductivityVSMeasurement precision

Solution Approach 1:

The system applies different evaluation strategies to different aspects of definition quality: rule-based checks for structural correctness, statistical analysis for conciseness metrics, graph theory algorithms for circularity detection, and language modeling for understandability assessment. Each local aspect receives specialized treatment appropriate to its nature, maintaining precision while enabling automated high-speed evaluation.

Inventive Principle:
Principle #3Local quality

Solution Approach 2:

The evaluation system combines multiple machine learning models and evaluation methods into a composite assessment framework. Different model types (rule-based, statistical, graph-based, language models) are integrated to evaluate various quality dimensions, creating a robust composite system that maintains high accuracy across diverse assessment criteria while operating at automated speeds.

Inventive Principle:
Principle #40Composite materials

Data Source

PatentEP4700604A1System for preparing machine learning training data for use in evaluation of term definition quality
Publication Date: 2026.02.25 COLLIBRA BELGIUM BV
  • EP4700604A1 patent drawingFigure 1
  • EP4700604A1 patent drawingFigure 2
  • EP4700604A1 patent drawingFigure 3

AI summary

A system for preparing machine learning training data for use in evaluation of term definition quality. The system can include a server having at least one server processor and at least one server memory for storing a plurality of terms with corresponding definitions, and a plurality of client devices each having at least one client memory device and at least one client processor. The client processor programmed to receive at least one of the plurality of terms and its corresponding definition from the server, display the term and its corresponding definition, and receive an indication of whether the definition satisfies one or more definition quality guidelines. The server memory includes instructions for causing the at least one server processor to receive the indications from the plurality of client devices and label each definition as satisfying each of the definition quality guidelines or not based on the received indications.